* copyright (C) 1984-2019 merrill consultants dallas texas usa
Yüklə 28,67 Mb.
|
|
extremely small amount of time that is not counted.
connection-reuse is in effect. I can't see any record said lacking of PK46171 will directly make qwacbjst to zero. - And this note on why QWACBJST can validly be zero. From dump, we can see CPUTIME is 0 but zIIP time is >0 and this is a zIIP eligible distributed SRB. Thus this is working as expected. The CP time can be 0 since all time might be on a zIIP at the time of the first clocking. As long as either the CP or the zIIP time > 0, that is normal. Thanks to Sieghart Seith, FIDUCIA IT AG, GERMANY Thanks to Derrick Haugan, MetLife, USA. Thanks to Lisa Ouellette, Wachovia, USA. Thanks to Jim Lazowski, NAV-INTERNATIONAL, USA. Thanks to Uncha, DB2 Level 2 Support, GERMANY. Thanks to Ronald Lobodzinski, DB2 Level 2 Support, USA.
IMACSPCK &MACSPCK is defined in VMXGINIT and referenced in the VMXGINIT IMACSPCK tailoring member, though unlikely to be used. Dec 5, 2007 Thanks to Chuck Hopf, Bank of America, USA.
JCLTEST8 were missing in the TESTOTHR/TESSOTHR steps, and there JCLTESS8 were inconsistencies in the TEST vs TESS members that JCLTEST9 were corrected in these four test members. The main JCLTESS9 purpose of these TEST/TESS jobs is to confirm that your Dec 4, 2007 "USERID.SOURCLIB" tailoring did not cause any errors in the TESTIBMx steps, so a failure in a subsequent step due to a missing DD statement should not prevent you from moving to JCLPDB8/JCLPDB9 testing. Thanks to Eric Barnes, Scottish and Southern Energy, SCOTLAND. Change 25.262 The 233 DDU files needed for ITRM sites to create ITRM TYPETNG table definitions for each of the 233 datasets built Dec 4, 2007 by MXG's TYPETNG (for CA's NSM product, formerly TNG). There is also the cpddudef.sas program that is used to generate all table definitions,. You will need to replace &YOUR_PDB_PATH & Your_DDU_PATH in cpududef.sas with your paths. The creation was run under a SAS ITRM interactive session. The itrmtng.sas file contains all 234 files in IEBUPDTE format; the individual files can be created by using the IEBUPDTE.SAS program in the MXG Source Library with the itrmtng.sas file as it's input. Thanks to Xiaobo Zhang, CheckFree, USA.
VMAC110 Defer Interval, the value specified in the site's LGDFINT Nov 29, 2007 parameter, but the MXG format only printed 2 decimals; the value is normally in milliseconds, so the format TIME13.3 is now used so a value of 5 ms will print as 00:00:00.005. Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.
VMAC30 for TYPETASK='OMVS' record; it is now protected twice, Dec 1, 2007 in the SUBSTEP loop, and at the OUTPUT statement. Thanks to Carl Sablon, KBC, BELGIUM. Change 25.259 VMXGALOC bumped to the next week's PDB one day early, and VMXGALOC could do even worse if the week-start-day was not Monday. Nov 29, 2007 The logic was revised for both errors by this CodeShark. Thanks to Patrick Holloman, Zions Bank Corporation, USA. Change 25.258 Intentionally Blank Change (a/k/a skipped). Change 25.257 Support for ICSF HCR7750 SMF Logging Update for TKE adds FORMATS these new variables to SMF type 82 subtype 16 TYPE8216: VMAC82 SMF82PAL='LENGTH OF*FIXED*AUDIT*DATA*' Nov 29, 2007 SMF82PDE='DESCRIPTION' SMF82PFI='FUNCTION*ID' SMF82PFR='FUNCTION*RETURN*CODE' SMF82PTA='TKE*AUTHORITY' SMF82PUS='USER ID*NONCE*TSN' and incorrectly spelled SMF82PDK is now SMF82PBK. New MG082RC format decodes SMF82PFR. Thanks to Greg Burt, Fifth Third Bank, USA.
TIMEBILD It's value is now the name of the member, suffixed with a UTILRMFI numeric value when multiple values are needed. Previously VMXGRMFI tests were for a simple numeric value, which triggered VMXGSUM unwanted debugging diagnostics from other code members. VMXGSUME And, UTILEXCL now exploits &MXGDEBUG with BEFORE/AFTER UTILEXCL location for each of the optional CICS data segments, so Nov 21, 2007 diagnosis of user tailoring errors will be faster! For example, you could use: OPTIONS FIRSTOBS=3800 OBS=3800; %LET MXGDEBUG=IMACEXCL; %INCLUDE SOURCLIB(TYPE110); if you had an error after IMACEXCL/IMACICxx members were tailored, and the error was in _N_=3800th record. Change 25.255 Support for Action Software's EventAction SMF User Record EXEVTA00 creates new datasets: EXEVTA01 DDDDDD DATASET DESCRIPTION EXEVTA02
EXEVTA03 EVTA00 EVTA00 DATASET CHANGE OR REFERENCE EXEVTA04 EVTA01 EVTA01 CHANGESMXSMF) EXEVTA05 EVTA02 EVTA02 REFERENCES MZSMF) EXEVTA06 EVTA03 EVTA03 CHANGE ACTION CONTROLS EXEVTA07 EVTA04 EVTA04 TEST ACTION C506) EXEVTA08 EVTA05 EVTA05 COMMAND CONTROL C507) EXEVTA09 EVTA06 EVTA06 CHG.DISTRIB TRANSMITS MZSMF) EXEVTA0A EVTA07 EVTA07 REF.TRACKING BY MEMBERS CS501) EXEVTA0B EVTA08 EVTA08 UPDATE TO EXCLUDES TABLE C41E) EXEVTA0C EVTA09 EVTA09 UPDATE TO BLACKOUT TABLE C427) EXEVTA0D EVTA0A EVTA0A UPDATE TO DATASET OPTIONS C404) EXEVTA0E EVTA0B EVTA0B UPDATE TO MEMBER OVERRIDE C405) EXEVTA0F EVTA0C EVTA0C OPTIONS AT OID LEVEL C40F) EXEVTA10 EVTA0D EVTA0D GLOBAL PARMS C401) EXEVTA11 EVTA0E EVTA0E DATA SET GLOBAL OPTIONS C401) EXEVTA12 EVTA0F EVTA0F PXC EXEVTA40 EVTA10 EVTA10 UPDATE TO USER GROUP EXEVTA50 EVTA11 EVTA11 EXCLUDES NOT REF.TRK) EXEVTA51 EVTA12 EVTA12 UPDATE TO CMD>TRACK OPTIONS EXEVTAF0 EVTA40 EVTA40 CHANGE REQUEST DELETE EXEVTAF1 EVTA50 EVTA50 USS CONTROLS IMACEVTA EVTA51 EVTA51 USS EXCLUDES TYPEEVTA EVTAF0 EVTAF0 ACCOUNTING RECORDS C50C) TYPSEVTA EVTAF1 EVTAF1 USS CHANGS AND REFERENCES VMACEVTA VMXGINIT
FORMATS Nov 18, 2007 Thanks to Craig Collins, State of Wisconsin DOA DET, USA.
ANALACTP UOW level keeping track of total response and CPU, Nov 18, 2007 the longest package, the first 10 packages. Jan 8, 2008 Jan 8: Typos in the untested code were discovered/fixed. Example expected UOWIDCHR variable had been added to your DB2ACCTP dataset, but didn't show how to, or note it could be removed from ANALACTP example. Thanks to Dan Almagro, Automobile Club of Southern California, USA. Change 25.253 Support for new NTSMF MSSQL Objects. EXNTQLBA DDDDDD DATASET DESCRIPTION EXNTQLBN EXNTQLBS NTQLBA MSQBROAC MSSQL:BROKER ACTIVATION EXNTQLBT NTQLBN MSQBUFND MSSQL:BUFFER NODE EXNTQLCA NTQLBS MSQBROST MSSQL:BROKER STATISTICS EXNTQLCL NTQLBT MSQBRODT MSSQL:BROKER/DBM TRANSPORT EXNTQLCT NTQLCL MSQCLR MSSQL:CLR EXNTQLCU NTQLCA MSQCATME MSSQL:CATALOG METADATA EXNTQLES NTQLCU MSQCURMG MSSQL:CURSOR MANAGER TOTAL EXNTQLPC NTQLCT MSQCURTY MSSQL:CURSOR MANAGER BY TYPE EXNTQLSR NTQLES MSQEXECS MSSQL:EXEC STATISTICS EXNTQLTR NTQLPC MSQPLANC MSSQL:PLAN CACHE EXNTQLWS NTQLSR MSQSQLER MSSQL:SQL ERRORS IMACNTSM NTQLTR MSQTRANS MSSQL:TRANSACTIONS VMACNTSM NTQLWS MSQWAITS MSSQL:WAIT STATISTICS VMXGINIT Nov 18, 2007 Thanks to Bob Gauthier, Albertsons, USA.
CONFIGW2 -MXGWPSV2. JCL procedure updated for WPS under z/OS MXGWPSV2 -VMXGINIT. Test for identification of WPS revised, code VMXGINIT was relocated to after TAPENGN was set for SAS: VMXGPRAL %IF %SYSPROD(WPS) EQ 1 %THEN %DO; ANALDB2K %LET WPSVER=&SYSVER; ANALHTML %LET SASVER=8; ANALMQMC %LET TAPENGN=WPDSEQ; ASUM42DS %END; ASUMCACH -CONFIGW2. CONFIG options now specify SEQENGINE=WPDSEQ. ASUMHSM -VMXGPRAL. Tests for ENGINE adds WPDSEQ to list of seqs. ASUMTALO Unrelated, SUM was added to PROC MEANS output. CICSTRAN -VMXGINIT. WPS does not yet support VIEWS; all members DB2PDB with /VIEW=XXXXXX were replaced with &VWxxxxxx GRAFRAID macro variables that are %LET to the correct JCLUOWP View-NAME under SAS but blanked under WPS. JCLUOWV This change will be reversed when WPS has UTILRMFI added support for Views. UTILUOW -ANALDB2K thru VMXGUOWT listed at left were "View-Revixed". VMXGCAPT WPS Level Tested successfully was Build (8460). VMXGSUM VMXGSUME MXG Newsletter FIFTY-ONE, VI.A WPS Technical Note reports VMXGUOTT 1. Current status of MXG Testing under WPS Betas Nov 2007. 1.j. MXG Support Position for testing of WPS Release. VMXGUOW 2. Run time comparisons. VMXGUOWT 3. Revision to SAS Clones article in MXG Newsletter FIFTY. Nov 19, 2007 4. Summary and statistics Jan 30, 2008 Jan 30: typo VMUM corrected to VWUM. Thanks to Brian Carter, EDS, UK.
VMAC7072 in their labels, but those are all '1024-BIT' counts and Nov 17, 2007 durations; their labels are now corrected. Thanks to Miguel F. Monferrer Carvajal, SPAIN. Change 25.250 Variable TARCELAP is now FORMATed TIME12.2. VMACTMNT
Nov 17, 2007 Thanks to Chuck Hopf, Bank of America, USA. Change 25.249 Variable TAUSRDAT was INPUT as $EBCDIC32. but can have up VMACTMO2 to 240 bytes of data; INPUT statement was revised to use Nov 17, 2007 TAUSRLEN to determine the length of user data input. Thanks to Rodger Foreman, Acxiom, USA. Change 25.248 The current VMXGSUM creates output variables with stored VMXGINIT LENGTH of 5 (z/OS) or 6 (ASCII), based on the value of VMXGSUM &MXGLEN (set in VMXGINIT), unless they are used in the Nov 18, 2007 SUMLONG=, MAXTIME=, OR MINTIME= arguments, which always create LENGTH 8 variables. You could change those lengths with an explicit LENGTH statement in OUTCODE=, or you could change the &MXGLEN value, but that would also change subsequent LENGTHs of all defaulting variables in subsequent steps in the same SAS session/step. The SUMBY= and ID variables are output in the same length they were in the input dataset, or in the INCODE= code if that is where they were created. They could be changes in the ORDER= argument. This change creates macro variable &LNSUMOUT which will only apply to VMXGSUM and makes all variables on which we do mathematical operations to be LENGTH 8. The default value of LNSUMOUT is blank, so the variables will have the original (shorter) length unless you set %LET LNSUMOUT=8; in your //SYSIN stream. Apr 2008: DO NOT USE LNSUMOUT. See Change 26.065. Thanks to MP Welch, SPRINT, USA.
VMAC120 caused INPUT STATEMENT EXCEEDED RECORD or INVALID DO LOOP Nov 13, 2007 CONTROL error; only SM120SNT=1 records had been read and this condition exposed an MXG logic error, now corrected. Thanks to Bjorn Helgestad, VPS ASA, NORWAY.
ASMRMFV to burp and die with an 0C4; this revision protects. Nov 13, 2007 Additional enhancements are noted in the changes in the Nov 17, 2007 ASM source comments. Thanks to Jon Whitcomb, Great Lakes Educational Loan Service, USA. Thanks to Jerry Urbaniak, Acxiom, USA. Change 25.245 CA PDSMAN product's SMF record created megabytes of data VMACPDSM when diagnostic trace records (LGRTYPE='D') were enabled, Nov 12, 2007 and CA's recommendation was to suppress them and also the Resource Monitoring (LGRTYPE='M') record processing. Thanks to Sudie Wulfert-Shcickendanz, Anheuser-Busch, USA.
VMACRACF two different lengths, 540 and 549, but MXG expected the Nov 12, 2007 549 length record, which caused INPUT RECORD EXCEEDED error when the shorter record was read. Both lengths are now supported. Thanks to Sean Angley, IBM, CANADA Change 25.243 The automatic PROC DELETE of the WORK.UNKNOWN dataset is TYPETNG removed, so that that dataset will exist after TYPETNG or TYPSTNG TYPSTNG program is used to process CA NSM (old TNG) data. Nov 6, 2007 If there are non-zero observations in WORK.UNKNOWN, it is very possible that some or all data will not have been output by MXG logic, so leaving WORK.UKNOWN will allow it to be tested for possible unknown data records. Thanks to Xiaobo Zhang, CheckFree, USA. Change 25.242 NDM record 'UC' is now output in NDMAE dataset. VMACNDM
Nov 5, 2007 Change 25.241 Support for CICS Transaction Gateway 7.1.0 new SMF 111 EX111CM statistics record: EX111CM datasets: EX111CS DDDDDD MXG MXG EX111CXE DATASET DATASET DATASET EX111CXI SUFFIX NAME LABEL EX111GD EX111PH 111CM T111CM CICS CTG COMMUNICATIONS MANAGER EX111SE 111CS T111CX CICS CTG CICS SERVER EX111WT 111CXE T111CXE CICS CTG EXCI SERVER INSTANCE IMAC111 111CXI T111CXI CICS CTG IPIC SERVER INSTANCE VMAC111 111GD T111GD CICS CTG GATEWAY DAEMON VMXGINIT 111PH T111PH CICS CTG PROTOCOL HANDLER Nov 5, 2007 111SE T111SE CICS CTG SYSTEM ENVIRONMENT Nov 21, 2007 111WT T111WT CICS CTG WORKER THREADS
EXITCICS MXG incorrectly believed that '20'x bit in MCTMNOPN was VMAC110 true when CICS/TS 3.2 SMF 110-1 records are internally VMAC112 compressed, but that bit only indicates that the compress Nov 3, 2007 option was enabled for this CICS region. This caused MXG Nov 13, 2007 to falsely report the EXITCICS decompression exit was not correctly installed, when Dictionary Records (MNSEGCL=1), which are never compressed, were read. MXG now tests for non-zero MCTSSCRL, which is the documented condition for a compressed CICS SMF 110 or 112 record. -VMAC112 was similarly changed to test for non-zero OMSPCL to detect compressed SMF 112 records. -This incorrect assumption had also been passed in my spec for the EXITCICS logic, which had just turned off that '20'x bit in its decompressed output. Now, EXITCICS sets MCTSSCRL or OMSPCL to zero after decompression. -If you previously assembled EXITCICS prior to this change you must reassemble with this revised EXITCICS member AND use the revised MXG's VMAC to read compressed records.
EXTPMALG JXSLMCC ='JXSLM*CONTROL*CENTER' EXTPMDBS JXSLMCC ='JXSLM*CONTROL*CENTER' EXTPMSLM -New SLM JOB Statistics subtype 5 creates TPMSLM dataset. FORMATS -New DBS POOL subtype 240 creates two new datasets: IMACTPMX TPMALG - Algorithm data VMACTPMX TPMDBS - DBS Pool data VMXGINIT -Jan 4: Variables JESNR JCTJOBID added to TPMSLM dataset. Nov 3, 2007 -Jan 9: GA records have corrected ETP truncation that was Jan 4, 2008 originally reported here. Jan 9, 2008 Thanks to Scott Barry, SBBWorks, Inc, USA. Change 25.238 Optional OMEGAMON BSC segment for CICS/TS 3.2 did NOT IMACICOB increase time-duration/count fields to the full 12-byte IMACICOC resolution that I had ASS-U-Med, which caused MXG ERROR IMACICO2 message INVALID STRTTIME. I made the same assumption for Nov 2, 2007 all three Omegamon segments with time/count fields, so I now assumed that the other two segments also have 8-byte fields (IMACICOB for OMEGDB2, IMACICO2 for OMEGAMON), so those two members also are reverted. And, of course, if Omegamon does increase their fields to full 12-bytes, yet another MXG change will be required. Thanks to Ray Dunn, CIGNA, USA.
VMAC7072 was not kept after MXG 23.09, but the same named variable Nov 1, 2007 LCPUWAIT='LPAR*WAIT*COMPLETION?' in PDB.TYPE70PR was, and that was the source of the problem. Now, a rename fixes this error, which was introduced in the infamous SPLIT70 rewrite.
Thanks to Enzo Rossi, Demand Technology Software, ITALY. Change 25.236 Change 24.141 with z/OS at 1.7 or earlier caused TYPE78IO VMAC78 dataset to have zero observations; MXG tested SMF78RSQ Oct 31, 2007 for zero or one, but SMF78RSQ does not exist when the Nov 1, 2007 RMF product segment is only 104 bytes. The test was revised to output TYPE78 for missing value, zero or one. But then, the duplicate observations created were NOT removed by the NODUP option, because the SMFTIME in the second or third replicates was not exactly the same value as the first, so the _STY78IO sort macro was rewritten to remove those with identical SMFTIMEs, and an extra DATA step is used to keep only the FIRST.SMFTIME instance. (The additional logic is invoked, but not needed, when the SPLIT 78 records have a valid SMF78RSQ value.) Thanks to Peter B. Hopper, CSC, AUSTRALIA. Thanks to Steven Olmstead, Northwestern Mutual, USA.
EXTSO030 new variables in existing Solaris MIB-2. EXTAI027 -Support for two new AIX Objects. EXTAI028 -Support for 10 new RedHeat Objects, many new Metrics EXTRH020 for existing RedHat Objects. EXTRH021
EXTRH022 EXTRH023
EXTRH024 EXTRH025
EXTRH026 EXTRH027
EXTRH028 EXTRH029
FORMATS VMACTNG
VMXGINIT Oct 31, 2007 Thanks to Xiaobo Zhang, CheckFree, USA.
VMAC85 by z/OS 1.7 are now input and kept in TYPE85ST dataset: Oct 28, 2007 R85B2ODK='BACKUP2*BYTES*DELETED FROF*OPTICAL' Nov 1, 2007 R85B2ODO='BACKUP2*OBJECTS*DELETED FROM*OPTICAL' R85B2ORK='BACKUP2*BYTES*READ FROM*OPTICAL' R85B2ORO='BACKUP2*OBJECTS*READ FROM*OPTICAL' R85B2OWK='BACKUP2*BYTES*WRITTEN TO*OPTICAL' R85B2OWO='BACKUP2*OBJECTS*WRITTEN TO*OPTICAL' R85B2TDK='BACKUP2*BYTES*DELETED FROF*TAPE' R85B2TDO='BACKUP2*OBJECTS*DELETED FROM*TAPE' R85B2TRK='BACKUP2*BYTES*READ FROM*TAPE' R85B2TRO='BACKUP2*OBJECTS*READ FROM*TAPE' R85B2TWK='BACKUP2*BYTES*WRITTEN TO*TAPE' R85B2TWO='BACKUP2*OBJECTS*WRITTEN TO*TAPE' R85NTE ='TAPE*VOLUMES*EXPIRED' R85RCLD ='RECALLED*OBJECTS*PROCESSED*THIS CYCLE' R85RCLK ='BYTES OF*RECALLED*OBJECTES*THIS CYCLE' and these variables added by z/OS 1.8 are input/kept: R85LOBD ='ROWS*DELETED*FROM LOB*STRUCTURE' R85LOBI ='ROWS*INSERTED*INTO LOB*STRUCTURE' Thanks to Harald Seifert, HUK-COBURG, GERMANY.
VMAC110 dictionary record, but the MNSEGCL flag that identifies Oct 27, 2007 the record IS a dictionary record was not printed.
VMXGINIT but it forced WPSVER=2; now, the actual WPSVER is stored Oct 26, 2007 in &WPSVER.
VMXGINIT the location in VMXGINIT was inside a DO GROUP that was Oct 26, 2007 only executed for SAS V8.2, causing UNRESOLVED MACRO when MXG executed under SAS Version 8.1. The statement was relocated so it is always executed, no matter what SAS version is used. Thanks to John Compton, ACS, USA.
VMACCIMS DBD segments when it was non-zero, but PTF BQI0129 for Oct 23, 2007 IMF 4.2 populated that previously reserved field, which caused INPUT EXCEEDED error and this error message: INVALID IMS TRANSACTION RECORD LENGTH=836 WITH xxx 48-BYTE DBDS EXPECTED AFTER COL=32765 _N_=1 Now, MXG only uses the 4.3 offset to DBDs when the IMF version is 4.3 or greater. Thanks to Siegfried Trantes, IDG, GERMANY.
VMACNMON number of CPUnn records increase and decrease as AIX adds Oct 23, 2007 or subtracts "virtual" CPUs, and when a CPUnn goes away, Nov 2, 2007 NMON wrote a short record, which caused INPUT EXCEEDED error. Now, MXG detects and deletes these short records. Note that the PDB.NMONINTV dataset, in NRCPUS variable, has the number of "real" CPUs. However, in this case, the value of NRCPUS was always 6, even though there were CPUnn segments with CPU14 (i.e., there should have been NRCPUS=7, as there are 2 "virtual" CPUs for each "real". -Temporary variables WORD11-WORD24 were not set to LENGTH $128, so they took the SAS default of 8-bytes for CHARs, causing character variables stored from them to also have a stored length of 8 bytes. Now, all WORDnn are $128, and specific LENGTHs for kept variables are used where needed.
-Variables NRCPU, PID, and PPID are now numerics. -NMON data value 'nan' is 'Not a Number' and is stored in the data records, causing INVALID DATA messages until each instance is protected with double question marks!! Sometimes spelled NaN. Thanks to Mike Woelke, Boeing, USA. Change 25.228 Protection for invalid SMF 14 record that had NUCB=2 but VMAC1415 only one actual UCB segment. This record caused ERROR: Oct 17, 2007 INPUT STATEMENT EXCEEDED RECORD LENGTH. Protection will print error message for first three bad instances. Thanks to William Carrol, Grange Insurance, USA.
VMAC7072 a Service Class versus a Reporting Class observation. It Oct 16, 2007 was not kept previously because the SMF manual mentioned only service classes, but actual data shows TYPE72DL can contain both Service and Reporting Class observations. Thanks to Harald Seifert, HUK-COBURG, GERMANY Change 25.226 UPRINDOC will PROC PRINT the NAME and LABEL of variables UPRINDOC and is used to create the example output in the ADOCxxxx Oct 16, 2007 members, and it also PROC MEANS all numeric variables. It's been in MXG for years, but never documented. Change 25.225 RMF III variable ENCCPUT is labeled 'CP*ENGINE*CPU TIME' VMACRMFV now, because it is recalculated to remove zIIP CPU time: Oct 16, 2007 ENCCPUT=ENCCPUT-ENCSUPT; Yüklə 28,67 Mb. Dostları ilə paylaş:
|